HXL-Data-Science-file-formats

HDP Declarative Programming conventions, early draft

This is an draft. See https://github.com/EticaAI/HXL-Data-Science-file-formats/issues/16.


Trivia:



High level goals

TODO: draft this

Conteiner format

  1. The preferred container format is YAML (https://yaml.org/)
    1. Platforms specific features, like the functionality of loading native directly in HDP tools:
      1. Should be ignored in existing YAML files
      2. Should NOT be exported with HDP+YAML format, even if well behaved HDP+YAML parsers would likely to ignore.
  2. JSON, in special for compatibility with other tools or for machine-to-machine HDP exchange or for specific low level calls, can be used as additional conteiner format.

Exchange of commands with foreign interface

  1. The recommended way of exposing HDP functionality to low level data processing that already not native on HDP is both exporting and accepting JSON format, without the full semantics of HDP (like translations of vocabulary)
    1. [INFORMATIONAL] Ansible uses JSON for data exchange of plugins.

Conventions for identifiers (key names)

Tokens meaning on keys

# Quick example 1
- hsilo:           # hsilo is an verb, without additional tokens
    linguam: LAT   # linguam is also an verb, LAT is an value

Note: Remove \ from \{\{ and \}\} if this document is renderized with them.

The << >>, <<! !>>, , <<!! !!>>, , <<!!! !!!>>

Draft from v0.8.3

Issue: ! character is sometimes used to mean ‘encrypted’, so maybe we should choose another character

Collaborative work

See collaborative-work.md

Security considerations

See security-considerations.md